-U
ssINST.
rc
AUG 30 1940
lRRAIR(
RATE
OF
OF
HYDRATION
PROPYLENE
by
William A. Merritt
B.S., Massachusetts Institute of Technology
1939
Submitted in partial fulfillment of the requirements
for the degree of Master of Science
from the Massachusetts Institute of Technology
1940
Signature of Author......................
Department of Chemical Engineering, May 16,
Signature of Professor
in Charge of Research ....................
1940
.o..o..o.....o..
Signature of Chairman of Department
Committee on Graduate Students..................
_
-- t
0
The Graduate House
Mass. Institute of Technology
Cambridge, Massachusetts
May 10, 1940
Professor George W. Swett
Secretary of the Faculty
Massachusetts Institute of Technology
Cambridge, Massachusetts
Dear Professor Swett,
In partial fulfillment of the requirements for the degree of Master of Science, I submit to you
this report on an investigation of the rate of hydration
of propylene by sulfuric acid.
Yours very truly,
William A.
238395
Merritt
r
iEl
..
Acknowledgment
The author wishes to express his appreciation
for the guidance of Professor E. R.
Gilliland during
the conduction of this investigation.
TABLE
CONTETS
OF
I.
SUMMARY ...........................
II.
INTRODUCTION .........................
III.
PURPOSE.. ......
IV.
PROCEDURE.. ......
V.
RESULTS ........
VI.
DISCUSSION OF RESULTS..................
32
VII.
CONCLUSIONS.......................
.
40
VIII.
RECOMMENDATIONS ......................
41
IX.
APPENDIX.........................
42
.......
1
.
5
................
........
.........
..
20
...........
..
....
.
.....
1
27
.............
A. Calculation of Theoretical Per
Cent Conversion at Equilibrium
vs. Temperature for Different
Pressures .........................
43
B. Calibration Plots and Calculations
therefor.........................
45
C. Method of Analysis................
52
D.
Sample Calculations (Data of
Run III) .........................
54
E. Summarized Data...................
56
F. Details of Apparatus..............
60
G. Bibliography......................
64
r
INDEX
Figure I
TO
DIAGRAMS
AND
GRAPHS
Graph of Theoretical Per Cent
Conversion at Equilibrium vs.
Temperature for Different
....
Pressures .....................
18
II.
Diagram of Apparatus..............
23
III
Diagram of Apparatus for Generating and Liquefying Prophylene....
25
Graph of Alcohol Conaentrations
vs. Time (Run I)................
28
Graph of Alcohol Concentrations
vs. Time (Run II).............
..
29
Graph of Alcohol Concentrations
vs. Time (Run III).............
30
IV.
V.
VI.
VII.
Graph of Alcohol Concentrations
vs. Time (Run IV)..............
VIII
Curve of Alcohol Concentrations
vs. Time Calculated from Lewist
Data for Batch Operations......
36
LLX
Flowmeter Calibration Plot.....
47
XL-
Thermocouple Calibration Plot..
49
XI.
Thermocouple Deviation Plot....
50
XII.
Pressure Gage Deviation Plot...
51
I.
SUMWARY
The purpose of this thesis was to study the rate of
hydration of propylene under flow conditions, using 60% sulfuric acid, moderate pressures, and temperatures around 100'C.
The desirability of obtaining accurate data on this subject
is due to the fact that the established method of producing
isopropyl alcohol from absorption of propylene in approximately 80% sulfuric acid, dilution, and subsequent rectification
of the acid ester involves the expense of reconcentrating the
sulfuric acid and also the loss of a considerable amount of
the acid in the reconcentrating process due to the oxidation
of organic substances left in the dilute acid after rectification.
At ordinary pressures and temperatures the rate of
absorption of propylene in 60% sulfuric acid is negligible,
and temperatures above 10(? C.
polymerization.
cause oxidation by the acid and
If the weak acid is to be used, the only
alternative is to employ pressure, in order to obtain reasonable reaction rate.
This thesis work was performed at moderate pressures,
ranging from 100-250 lbs./sq.in. gage, and temperatures from
94-160 C.
Sulfuric acid about 60% in strength was charged to
a copper-lined pressure reactor, and propylene continuously
recirculated through the reactor, the gas from the reactor
being sent through a condenser, from which the condensed
alcohol ran into a trap.
Make-up propylene was supplied from
a gas holder, and the gas was recirculated via the compressor
which kept pressure on the reactor.
Samples for analysis
could be taken from the reactor and from the condensate trap,
and the rate of flow of propylene was measured by a flowmeter.
Complete details of the apparatus may be found under
PROCEDURE.
Four runs were obtained, the first two at 120 and
100 lbs./sq.in. gage respectively and about 100*C., and the
third and fourth at 150 and 250 lbs./sq.in. gage and 160'C.
The propylene flow rates for the four runs were respectively
9.50, 7.02, 4.76, and 6.03 cu.ft./hr. at S.T.P.
In spite of
the comparatively high rates of gas flow in the first and
second runs, practically no vapor condensate could be obtained
from the condenser trap, due to the fact that the operating
temperatures were too low.
The alcohol content of the acid
in the reactor at the end of the first two runs was 23.1% and
19.1% respectively, and in each case considerable organic
material was found as a separate layer floating on top of the
acid layer, although at the higher temperatures of Runs III
and IV more of the organic material was formed.
In the last
two runs the equilibrium alcohol content of the vapor from
the reactor was 15.2% and 18.2% respectively, the slightly
nigher concentration for Run IV apparently being due to the
higher operating pressure.
Theequilibrium alcohol concen-
trations of the acid in the reactor for Runs III and IV were
28.4% and 29.1% respectively, a substantial increase over
the corresponding concentrations for Runs I and II, which
were operated at lower pressures and temperatures.
At the higher temperatures of Runs III and IV a
considerable condensate rate was obtained, being 56.0 and
64.0 cc. per 15 minutes respectively 4t equilibrium).
From
the condensate rate data and the flow of propylene to the
reactor, the molal ratio of alcohol to propylene in the vapor
from the reactor could be calculated for Runs III and IV, and
also the per cent conversion at the dynamic equilibrium conditions, dynamic equilibrium conditions being defined as when
the rate of increase of alcohol concentration in the condensate and reactor acid approached zero.
The per cent conver-
sions so calculated for Runs III and IV were 9.4% and 10.1%.
The theoretical % conversions at equilibrium for these conditions as calculated from equilibrium and fugacity data are
16.1% and 17.3%.
In spite of the comparatively low vapor
rates employed in these runs it seems remarkable that the time
of contact attained between gas and acid was sufficient to enable the true equilibr-ium conditions to be approached so closely.
The faster flow rate in Run IV must be responsible for the
shorter time required to reach dynamic equilibrium conditions,
the higher alcohol concentration attained at equilibrium
apparently being due to the higher pressure.
From this data it seems that for commercial operation
probably the most advisable operating conditions would be not
less than 120'C.
and about 150 lbs./sq.in. pressure. IXtcreas-
ing the pressure would not appreciably increase the alcohol
concentration obtained, which would be about 15%.
As opposed
to the present process of diluting the acid ester mixture,
r
rectifying, and reconcentrating the diluted acid, the
operation would be carried out by rectifying the dilute
alcohol produced to the concentrated form, and adding water
at intervals to the acid to replenish that which has been
evolved as alcohol and water vapor.
At times the acid would
have to be changed, after the accumulation of too much
organic byproduct.
It is recommended that more data be obtained using
60% sulfuric and other pressures, temperatures, and flow
rates.
II.
INTRODUCTION
The problem of the utilization of synthetic petroleum
gas is becoming increasingly important because of the great
development of the vapor and liquid phase cracking inductries,
due to the demand for high anti-knock fuels.
Refineries are
overburdened with large quantities of gas of high olefin content, which at the present time are being burned as fuel, in
spite of the fact that they constitute a source of raw material
(6)
for the preparation of very valuable chemical products.
About 2,000,000 barrels of oil are cracked daily, producing over one billion cubic feet of gas of high olefin content.
The predominant olefin is the lowest one in the series,
ethylene, and a gas obtained as the result of a normal liquid
phase cracking process contains 12-14% ethylene, while gas from
vapor phase cracking contains 14-18% of that olefin.
The
(2)
propylene content is somewhat lower.
Typical analyees of the
gas from liquid phase and gas phase cracking units are as follows:
Methane and Hydrogen
Liquid Phase
Cracking (Cross)
Gas Phase
Cracking (Gyro)
46.3 %
35.4 %
1.2
22.9
16.6
13.3
5.3
18.1
16.2
0.0
Butenes and Butadiene
4.3
5.5
Butanes
6.1
1.4
Ethylene
Ethane
Propylene
Propane
Pentanes and heavier
4.0
oo100.0 %
3.4
100.0
.
~
The composition of the olefin fraction of a gas representing roughly 50% of a typical vapor phase cracking gas is in
the order of 55% ethylene and 30% propylene.
Thus the problem of utilizing this vast source of olefins
is commanding much attention at the present time, and most of the
work to date has been directed toward the production of alcohols
by a hydration process.
The most successful process in commercial operation
depends upon the absorption of the olefin gas in strong sulfuric
acid, resulting in the formation of the alkyl ester, which is then
(6)
hydrolyzed to yield the alcohol.
In more detail, the gas contain-
ing a mixture of ethylene, propylene, and other olefin and hydrocarbon vapors is first treated with caustic to remove traces of
hydrogen sulfide, and then with sulfuric acid of specific gravity
1.6, at a temperature of 300 C., to remove other olefin material,
permitting the propylene to pass unabsorbed.
In a second treat-
ment, the propylene is absorbed in sulfuric acid of specific
gravity 1.8, at a slightly higher temperature, and the solution of
isopropyl acidsulfate is allowed to build up.
The mixture is then
hydrolyzed with water, in which process partial conversion to the
alcohol takes place, and the diluted solution is then rectified,
completing the hydrolysis and yielding an overhead product of the
constant-boiling mixture 91% isopropyl alcohol by volume.
The
process is usually run continuously, the diluted sulfuric acid
being reconcentrated and used over again.
Many patents have been taken out relating to this process,
and much experimental work done to effect improvements in the
_ ~.
, -r~aL
u'-^IC,
olefin absorption, such as the addition of mineral oil to the
acid in order to stimulate absorption of the olefin prior to
actual esterification, the use of catalysts in the sulfuric
acid, and the use of higher temperatures and pressures in
order to increase the rate of reaction.
The cost involved due to the necessity of reconcentrating
the sulfuric acid akes the desirability of using a weaker sulfuric acid, such as 60%, apparent.
Also, acid is lost in the
reconcentrating process due to the oxidation of organic substances left in the dilute acid after rectification.
The rate of reaction with 60% sulfuric acid is negligible
at ordinary temperatures and pressures, and temperatures greater
than 100 C. tend to cause oxidation by the acid.
Several pro-
posals have been made to raise the pressure, however.
Some
actual data have been obtained by the Standard Oil Company of
(20)
New Jersey in several pressure tests. A gas containing 18-20%
propylene was shaken in a bomb with 60% sulfuric acid at
400-600 lbs./sq.in. pressure and 80-100"C.
In two hours the
absorption amounted to 1.5-2.4 mols of propylene per mol of
acid.
In flow experiments, however, the results obtained were
not nearly as satisfactory.
A semi-plant tower about fifty
feet high, designed to provide a long time of contact between
gas and acid, when operated on a gas containing 18% propylene
gave a reduction only to 10-15% in the gas issuing from the
top.
In another flow experiment, a one-plate column was
constructed with a twenty-foot acid depth.
At entering gas
rates of 30 cu.ft./min., the propylene content fell from
~ . lar~e "~ere
18% to 15%,
while at entering rates of 150 cu.ft./min. the
propylene content fell to 10%, presumably due to the fact
that at the high gas rate a large quantity of small bubbles
was produced rather than a few large ones, thus providing
better gas to acid contact.
In other words, while excellent
results were obtained in the case of the batch experiments,
for some reason the absorption was not nearly as satisfactory
in the flow experiments.
(20)
It was the purpose of C. Lewist thesis to obtain more
data on the mechanism of absorption under pressure, using a
batch reactor.
He made one three-hour run at a temperature
of 100'C., pressure 500 lbs./sq.in., and using 60% sulfuric
acid and pure propylene gas.
Besides the acid layer, which
contained dissolved propylene, isopropyl alcohol, the easily
hydrolyzable mono-sulfate and the difficultly hydrolyzable
disulfate, Lewis obtained 60 cc. of a yellow liquid insoluble
in the acid layer and floating upon it, constituting side
reaction products due to oxidation and polymerization.
Taking liquid samples periodically he found that at the end
of two hours the concentration of total (free and combined
as ester) alcohol leveled off at about two mols of alcohol
per mol of acid, rising to 2.1 within the next hour.
From
a curve of concentration vs. time, which rose in practically
a straight line for the first
two hours,
he concluded that
the mechanism was not one of rapid formation of ester at the
surface followed by diffusion through the liquid, that is,
that the rate of absorption is proportional to the concentration
~ -n~a9N 'Ira_
~
_- L
at saturation minus the instantaneous concentration, since
if this were the case the concentration vs. time curve
would have fallen off gradually instead of rising in a
straight line.
Lewis concluded that as the alcohol content
of the acid increased, the solubility of the propylene was
maintained due to the solvent action of the alcohol.
As opposed to this process using strong sulfuric
acid and involving the intermediate formation of the alkyl
ester before the production of the alcohol, much work has
been done on the direct catalytic hydration of propylene and
ethylene using as catalysts activated metals and dilute
mineral acids.
(49)
Klever and Glaser reported that in the absence of
acid at 150 0 C. and 190 atmospheres pressure only 0.00011
mols of ethylene per mol of water were hydratedafter 17 hours,
and at 2000 C. and 100 atmospheres only 0.0008 mols reacted,
while for the same length of time but in the presence of
1.93% HCI more than a mol of ethylene was hydrated per mol
of acid.
They also state that even with two or three %
mineral acid the reaction rate is too slow to be of commercial
importance.
(5)
Apparently the first patent on this phase of the work
was granted in 1924 claiming the production of alcohols when
olefins were heated with water at 150 -250 C. under pressures
from 20-200 atmospheres.
The water should preferably be
acidulated, and the reaction assisted by "surface action.9
(17)
A patent granted to Johannsen and Gross in this country
in 1927 specified pressures above 20 atmospheres and tempera-
~ ..
~~bY4
CLIICb-
~
_~e
10.
tures above the boiling point of water.
the body of the patent it
is
In an example in
claimed that by heating 2.0
grams of ethylene with 10 cc. of 4.3% HC1 at 220 0 C. a 63%
yield of ethard was obtained.
The use of hydrochloric acid,
sulfuric acid, and phosphoric acid as catalysts is specified.
Considerable work has been done using as catalysts
the salts of heavy metals having a slightly acid reaction,
and many "activated" metals, including gold, silver, copper,
iron, cobalt, nickel, chromium, tantalum, vanadium, molybdenum, manganese, and compounds of these such as the oxalates
and carbonyls.
Metal oxides have been used, such as the
oxides of aluminum, thorium, tilanium, tungsten, and chromium.
(51)
The work of Swann, Snow, and Keyes, in 1930, on the
hydration of olefins in the liquid phase is one of the most
complete in this field.
Using pressures from 600-800 lbs./
sq.in. at a temperature of 135°C., and employing a 5% HC1
catalyst, they showed that the highest concentration of
isopropyl alcohol obtainable in the liquid phase after one
hour was 0.63% by weight.
(14)
Gunness investigated the system ethylene, ethanol,
water, in the temperature range 174-3076C. and pressure range
82-264 atmospheres.
His object was to find an expression for
the free energy of hydration of ethylene in the gas phase
from the equilibrium data, employing the relation AF=--R'441.
The equilibrium constant was calculated from vapor phase data
at equilibrium, although had data been available on the
_i~
~..
. r~u~blrrs~-~sa_
11.
activity coefficients of the reactants in the liquid phase,
the liquid phase data might have been used, since both phases
were in equilibrium.
Gunness performed batch experiments,
using a 3 mol % sulfuric acid solution as catalyst, maintaining constant total pressure during each run, and providing
means of constant agitation of the gas and liquid phases.
When the concentration of the liquid phase reached a constant
value, a vapor sample was condensed and analyzed.
(24)
The thesis of Majewski was directed toward obtaining
expressions for free energy change in the hydration of propylene
in both phases, and also data on the rate of reaction, the
effect of temperature and pressure, and the nature of the side
reactions occurring.
Working in a pressure range 95-503 atm. and a temperature
range 160-3600 C., and using dilute phosphoric acid catalysts, he
studied the rate of reaction in the liquid phase by periodic
sampling.
The vapor at equilibrium was analyzed for isopropyl
alcohol, isopropyl ether, water, and polymer.
The principal
constituents analyzed for in the liquid were isopropyl aocohol
and dissolved propylene.
His results showed that the concentra-
tion of alcohol in bothphases increased as the temperature decreased and the pressure increased.
was found in the liquid phase.
No normal propyl alcohol
The range of alcohol concentra-
tions obtained in the liquid phase are:
at 2050C. and pressure
range 95-503 atm., 110-201 gms./liter, while in the vicinity of
280CC.
and the same pressure range, only 19.9-48.4 gms./liter.
Isopropyl ether, due to its high vapor pressure, was
found entirely in the vapor phase, and results indicated that
12.
lower temperatures were more conducive to its formation.
Below 2100C. no polymer was formed, but above this temperature it became a more predominant reaction product as the
temperature was raised.
At 2400C.
over one hour was re-
quired for its appearance, while at 290 C. 35 cc. were formed
in the same time.
The effect of pressure was to increase the,
amount of polymer formed and also its distillation temperature.
The densities and indices of the polymers were found
to resemble those of long-chain unsaturated hydrocarbons or
cyclics rather than complex alcohols or ethers.
At low temperatures a catalyst was found essential,
and it was believed that the reaction took place almost entirely in the liquid phase.
At least six hours were normally
required to reach equilibrium.
At low temperatures, when
side reactions such as oxidation and polymerization were
negligible, dilute sulfuric and hydrochloric acids were substituted and gave the same equilibrium conditions as did the
phosphoric acid.
The two concentrations of phosphoric acid
used, 7.7% and 12.1%,
gave the same equilibrium conditions
but the initial rate was twice as fast in the latter case.
In no run was any evidence found of the acid alkyl ester in
the liquid phase.
Majewski concluded that the effect of increasing the
pressure was to increase the solubility of the propylene in
its liquid phase and its concentration in the vapor phase, and
the net conversion to alcohol.
The conversion to isopropyl
ether varied as the second power of the propylene concentration,
and the conversion to polymer by some power of the propylene
~i~
nssrr--
..
'Lp-
13.
concentration greater than one.
Temperature affected the
equilibrium, the rate of reaction, and the nature of the side
reactions.
Increasing the temperature increased the polymer
formation.
The duration of the runs varied with the conditions
of operation, about 20 hours being required to reach equilibrium at 165'C., while at 2800 C. several hours were sufficient.
In the liquid phase, the principal reaction was the formation
of isopropyl alcohol, and no normal propyl alcohol was formed.
At temperatures from 160-165C.
and pressures from 95-503
atmospheres about 20 hours were required to attain equilibrium, but 45-65% of the reaction took place in the first two
hours, the reaction rate then falling off.
At the higher
pressures he observed a characteristic falling off of the
alcohol concentration with time after 3C hours, explained by
the formation of large amounts of side reaction products in
the vapor phase which decreased the partial pressure of the
propylene in the vapor and hence its solubility in the liquid,
causing the decomposition of the alcohol already formed.
At the higher pressures the time required to attain a
given liquid concentration was shorter and the maximum liquid
concentrations were greater than at the lower pressures.
At 200-210°C.
the concentrations attained were not so high
as at 160-165 0 C., but the reaction rates were so high that
95% of the equilibrium conversion was obtained in two hours,
the time required to reach complete equilibrium being about
14 hours.
L
In a comparison of the reaction rates produced by
m
-r~a4a-Ylp-
14.
various catalysts, Majewski tried 12.1% and 7.7% phosphoric
acid, 2.2% sulfuric acid, 1.5% hydrochloric acid, and pure
water.
The 12.1% phosphoric acid doubled the initial reaction
rate but did not change the final equilibrium concentration
attained by the 7.7% acid.
Research has not been confined entirely to ethylene
(25)
and propylene, however.
Marck and Flege
investigated the
catalytic vapor phase hydration of 2-butene at 427-528 0 C.
under conditions of continuous flow and pressures from
6000-5000 lbs./sq.in.
They obtained only low yields of
butanal, and concluded that as the molal ratio of steam to
butene is increased the conversion to the alcohol increases
and polymer formation decreases.
from 0000 to 5000 lbs./sq.in.
Increasing the pressure
increased polymerization to
a greater extent than hydration.
They-employed as catalysts
phosphoric acid and cuprous chloride supported on pumice.
(30)
Stanley, Youell, and Dymock have determined equations
for constants for the hydration reactions of ethylene and
propylene as functions of temperature.
They employed a flow
method at atmospheric pressure, working in the temperature
range 150-250 C. and using acid phosphates containing manganese and boron as catalysts.
Equilibrium was approached
from both the synthesis and decomposition sides.
gas was passed first
through a wet test meter,
The olefin
then a
saturator, which in the case of the synthesis experiments
contained pure water, and in the decomposition runs contained
_~
--1
F
15.
5% aqueous alcohol, and then through a heated tube into
the catalyst tube, which was of pyrex and 600 cc. in volume.
The gases issuing from-the reactor were condensed by an ice
and salt cooling medium, and the excess olefin gas measured
in a wet test meter and analyzed in a Bone-Wheeler apparatus.
In the experiments with ethylene, gas rates varied
from one liter per hour at 1450 C. to 20 liters per hour at
2500C., over 500 cc. of the catalyst.
Employing a mean value
of the equilibrium constants obtained from the synthesis and
decomposition runs, the equation derived for the constant as
a function of temperature was, in the case of ethylene,
X
logKp =
-
7-
-
./f
where T is
absolute Kelvin,
and Kp is
the ratio of the partial pressure of ethyl alcohol to the
product of the partial pressures of water and ethylene at
equilibrium.
Since the equilibrium constant decreases with
increasing temperature,
the hydration reaction is
ei.er*ic,
the heat of reaction at constant pressure being 9600 calories
per mol.
In the experiments with propylene, the gas rates
varied from 4.0 to 8.9 liters per hour per 500 cc. of catalyst,
and the equation derived was log Kp
L
i60 -
6
oo0.
The corresponding heat of hydration is about 9000 calories
per mol.
The dilute aqueous solutions of isopropyl alcohol
produced in these experiments were analyzed by the method
of Cassar (Ind. Eng. Chem. 1927, 19, 1061).
Quoting from an early paragraph in their article of
July 1934 in the Journal of the Society of Chemical Industry:
c:
16.
RWith the exception of a fevw isolated determinations at somewhat elevated temperatures, little
is known about the influence of temperature on
the position of equilibrium in any of the reactions between olefin and water.
data is
This lack of
no doubt primarily due to the fact that
no really active catalysts have been found for
the hydration of olefins, and accordingly it has
so far not been possible to approach equilibrium
except at high temperatures, where side reactions
may become appreciable.
In the course of a pro-
longed study of a large number of possible catalysts
for the hydration of olefins, we have been fortunate in discovering active catalysts in the
presence of which equilibrium may be approached
at temperatures as low as 150 0 C."
The catalysts developed by them were acid phosphates
containing manganese and boron, prepared by evaporating down
manganese carbonate and boric anhydride in 2:1 molal ratio
with a definite weight of phosphoric acid in excess water.
The resulting solid was baked at 200-2500C.,
pressed into tablets.
The composition of this catalyst was
represented by the formula
between 2.4 and 0.6.
145 C.,
powdered, and
MnO,BO,xH,PO% where X varied
For the experiments with ethylene at
the catalyst consisted of granules of pumice impreg-
nated with 66% sulfuric acid containing about 5% silver sulfate.
1
-rr~srrru~a
17.
Both from a commercial standpoint and as a logical
sequence to what has been done previously, it seemed appropriate to study the rate of hydration of propylene in a flow
(20)
process, at low pressures, using 60 sulfuric acid. Lewis
studied the mechanism of batch absorption of propylene in
60% sulfuric acid, but before this thesis no work had been
done on the flow process at low pressures with the exception
of the tower experiments of the Standard Oil Company of New
Jersey. In this thesis the 60% acid was charged to a reactor,
and propylene gas recirculated through the reactor via the
compressor which maintained the pressure on the system.
The
vapor issuing from the reactor passed through a condenser,
from which the condensed dilute alcohol ran into a trap and
the excess olefin gas recycled, makeup propylene being
supplied continuously from a gas holder which floated on the
low pressure side of the system.
The temperature of the re-
actor ;was controlled electrically, and the flow rate of the
propylene measured by a flowmeter.
Samples could be with--
drawn from the reactor and condenser trap and analyzed for
alcohol content.
When the rate of increase of alcohol con-
centration in the reactor acid and condensed vapor became
practically zero, dynamic equilibrium conditions were considered to have been reached, and from compositions and flow
rates the per cent conversion of the propylene being sent
through the reactor could be calculated and compared with the
theoretical per cent conversion at equilibrium for the
operating temperature and pressure as calculated from the
------
- -- r--
-
----
--
iT4L'Z
i-:
--
-Zf+---
-
-
-.
hZOO
::
~"~t
r
--
-
0
-
.p-- ..bw/Pesur
-
17
--T
~
-
_~...:t...
,--
iri.7~."
t
...
~~11 rrrut~L19.
(30)
equilibrium data of Stanley, Youell, and Dymock and the
fugacity chart for hydrocarbon vapors.
Figure I gives curves
of theoretical per cent conversion at equilibrium versus
temperature for different pressures,
and it is apparent
that for a given temperature increasing the total pressure
increases the per cent conversion obtainable at equilibrium,
but not by a great deal, particularly at the lower temperatures.
The calculations involved in plotting the curves of
Figure I are given in (A) of the APPENDIX.
Since high
temperatures are conducive to polymer formation, it was
desirable to carry out this investigation at temperatures
just high enough to provide a steady rate of alcohol evolution
from the acid in
the reactor.
~2-~dCbS
rr-L'20.
III.
PURPOSE
The purpose of this thesis is
to provide some
accurate data on the per cent conversion obtainable in a
flow process for the production of isopropyl alcohol
using 60% sulfuric acid, pressures from 100-250 lbs./sq.in.,
and temperatures from 100-150oC.
As originally planned,
the temperature was not to be carried above 1000C., but in
order to obtain a steady evolution of alcohol from the
acid in the reactor it was found necessary to raise the
operating temperature to 160'C.,
with a corresponding in-
crease in k yme'r formation.
It was planned to make a series of runs at different
flow rates for each of several pressures, and compare the
periods of time required to reach dynamic equilibrium, and
also to compare the per cent conversions obtained at dynamic
equilibrium with the theoretical per cent conversions at
equilibrium for the given temperatures and pressures as read
from the calculated curves of Figure I.
..
7L'~LILllb~ b-
21.
IV.
Description of Apparatus:
matically in Figure II.
given in
(F)
PROCEDURE
The apparatus is shown diagramDetails of the apparatus are
of the APPENDIX.
Essentially the apparatus consisted of a closed
system, propylene being recirculated through the acid in
the reactor via the compressor which held pressure on the
system, and make-up propylene being supplied from a storage
bottle maintained under a head of water on the low pressure
side of the system.
Samples of the acid from the reactor could be withdrawn for analysis, and samples of the condensed vapor from
the reactor could be withdrawn from the trap beneath the
condensor.
Operating Technique:
Before starting a run the acid was
charged to the reactor, and the reactor slowly brought up
to temperature.
The quantity of acid charged was 120-150 cc.,
and it was introduced through the sampling valve at the
bottom of the reactor by means of a funnel which served as
a level indicator.
The time required to bring the reactor
up to temperature was usually about fifteen minutes.
The
control valve was then closed and propylene flushed through
the system by allowing a slight pressure to build up and
discharging the air-propylene mixture through the liquid
sampling valve at the bottom of the trap.
After several
_
rrcac '--
22.
minutes of flushing it was considered that at least
eighty per cent of the air had been expelled and the
run was started.
After the pressure had built up almost to the
desired operating pressure, the control valve was cracked
and care was taken not to permit the static pressure as
indicated by the flowmeter static pressure gauge to rise
above the mercury equivalent of the four-foot water head
on the supply of propylene.
A satisfactory flow rate was
normally obtained without building up too great a static
pressure on the low pressure side of the system.
By
regulating a bypass valve on the compressor in conjunction
with the pressure control valve the flow rate could be
varied independently of the pressure on the reactor.
The propylene in the supply bottle was replenished
periodically, and at times it was found convenient to leave
the cylinder-valve cracked and continually bubble propylene
into the storage bottle rather than fill it at intervals.
Samples were taken about every hour from the reactor and condenser trap, and occasional samples from the
bubbler, although this was given up after it was apparent
that the acid and alcohol concentrations were negligible.
The acid and condensed vapor samples were analyzed for
alcohol by the Pomidorf method of analysis as modified by
Gunness (see Appendix), the acid samples being diluted with
twice their volume of water and hydrolyzed by refluxing
FI&vee--=
DIA6 AAAT OFA)P,94RATU.5
lo wme te Y,
_
rr~ea~--c-~lk
24.
before being analyzed.
Dynamic equilibrium conditions
were considered reached when the alcohol content of the
condensed vapor from the reactor failed to rise appreciably
over a period of one-half hour. The rate of production of
alcohol at dynamic equilibrium was found by allowing the
condensate to accumulate for 15 minutes at the end of the
run.
The motor was then stopped, the §ressure released
from the system, and the acid in the reactor withdrawn and
examined for polymer.
noted.
The volume of acid and polymer were
During the run, normally about two times per run,
the water in the reactor was replenished, and an attempt
was made to keep the concentration of the acid at about 60%.
The pressure was released each time this occurred and the
make-up water introduced through the sampling valve at the
bottom of the reactor.
Propylene Generator.
Figure III is a diagrammatic sketch of
the apparatus used to produce and liquefy the propylene
used in the runs.
The reactor contained alumina, and the
propylene was produced by introducing liquid isopropyl
alcohol (91%) at the bottom of the heated reactor and allowin g it to be dehydrated, the propylene-water vapor mixture
being passed through a condenser which condensed out the
water, the propylene being stored in a gas holder.
The
temperature of the catalyst was maintained at 300-350 C.,
and it was separated from the iron wall of the reactor by
I1
%-
-
,.
-
-
-
co vn4ewier
V-aOor
Ao/S r
&-te-Cr
3AL
e
Coen
c.y /
"de
r~rurl~~'Pa
26.
a layer of asbestos, since contact of the gas with the
metal promotes kolymerization of the propylene.
The
capacity of the gas holder was 65-40 cu.ft., and since
this represents about 4.2 lbs. of propylene one full gas
holder was considered sufficient for the runs, particularly in view of the fact that several pounds were already
available in a discarded cylinder.
The propylene was liquefied in a high pressure
cylinder, by combined compression and cooling.
The cooling
medium was dry ice in gasoline, and the compressor used
during the runs was used to keep pressure on the sylinder.
The gas was fed from the holder through calcium chloride
drying tubes, then through the compressor, and before entering the cylinder it was cooled partially by passing
through a coil immersed in an ice-salt bath.
At the tem-
perature attained by the dry ice the vapor pressure of the
liquid propylene is about 40-50 lbs./sq.in., so when the
pressure would slowly rise to about 80 lbs./sq.in. it was
apparent that air was being compressed and by means of a
pop-off valve the air and other inerts were periodically
removed during the liquefaction.
When the gas holder was
emptied the cylinder was allowed to come up to room temperature, during which time the pressure of the propylene
rose to about 180 lbs./sq.in., the existing vapor pressure.
The cylinder was then connected to the five-gallon bottle
as described above.
-- -
_
27.
V.
RESULTS
The results of this thesis are presented in
the following table and Figures IV, V, VI, and VII,
which are plots of alcohol concentration vs. time for
the four runs.
RUN I RUN II
Pressure (lbs./sq.in.gage) ......
120
Flow rate (cu.ft./hr.at S.T.P.).
Temperature of reactor IF
RUN III RUN IV
150
250
9.50 7.02
4.76
6.03
200
94
220
106
320
160
320
160
60
60
60
60
5.75
4.17
15.2
18.2
Alcohol concentration in reactor
acid at equilibrium (wgt.%)... 23.1 19.1
(excluding organic layer)
28.4
29.1
Alcohol production at equilibrium (mg ./hr. ) ............
(mmols/br.). .... . ......
---
34000
568
46200
771
--
6020
7630
--
0.104
0.112
Per cent conversion at equilibrium................ ...... ..
9.40
10.13
Theoretical per cent conversion
at equilibrium calculated
from equilibrium and fugacity
date ............
............
16.1
17.3
....
Per cent sulfuric acid used.....
Time re uired to reach equilibrium (hrs.)...................
100
5.00 4.67
Alcohol concentration in condensed vapor at equilibrium
)......................
(wgt.
Propylene flow rate (mmols/hr.)..
alcohol
Molal ratio proplene) in vapor
from reactor at equilibrium...
F-7cr um
AColos.
O
gr-To
i-)c
j
L 30
Qt
AI* rec.
P
/
3
-7-elpc (A4-S)
s
Al-co#OA- CON CSiV 7-RATIfp,'S
Oeu .W
4a
it
*t
ocM't~y~r
~C44~V~
"21 we.
e
,fi.r." trv ,b
0
7Z~e (Ars)
-r~bSci-~L
e
320
VI.
DISCUSSION OF RESULTS
The most important single conclusion that can be
drawn from the results of this thesis is that the conversion of propylene to isopropyl alcohol obtainable in
a flow process employing 60% sulfuric acid at temperatures from 100-1600 C.,
pressures from 100-250 lbs./sq.in.,
and the flow rates employed is a substantial proportion
of the theoretically possible conversion at equilibrium.
No figures for per cent conversion could be calculated in Runs I and II due to the fact that steady
evolution of alcohol from the reactor acid could not be
0
obtained at the operating temperatures of 940 and 106 C.
In Runs III and IV, however, operating at temperatures of
16 0 C., the per cent conversions obtained were 9.4% and
10.1% vs. the theoretical conversions of 16.1% and 17.3%
respectively, indicating that conversions definitely in
the range of those theoretically possible can be attained.
Furthermore, from an inspection of the curves for Runs III
and IV (Figures VI and VII) it is apparent that the
alcohol concentration curves might have risen a few per
cent higher in each case before becoming absolutely
horizontal,
that is,
dynamic equilibrium might not have
quite been completely reached.
The data of Runs III andI. indicate that at higher
pressures, 250 lbs./sq.in, as compared with 150 lbs./sq.in.,
the per cent conversion obtainable at dynamic equilibrium
33.
is slightly higher than at lower pressures.
and IV were both run at the same temperature,
Runs III
160 C.,
but the alcohol concentrations obtained in the condensed
vapor from the reactor in the two cases, 15.2% and 18.2%
were 3.0% apart, or a per cent difference of about 20%.
In Run IV the alcohol concentration obtained in the reactor acid was also slightly higher than in Run III, the
per cents being 29.1 and 28.4 respectively.
The effect
of flow rate is also brought out in Runs IIIand IV.
The
flow of propylene in Run III was 4.76 cu.ft./hr. as compared with 6.03 cu.ft./hr. in Run IV, and in the latter
case equilibrium conditions were reached in 4.17 hrs.
compared with 5.75 hrs. in Run III, shortening the time
required at 27.5% for a 26.8% increase in the rate of
gas flow.
Also, the volumetric production of alcohol at
equilibrium is higher at the higher velocity of Run IV
than in Run II.
Comparing Runs III and IV with Runs I and II, it
is apparent that increasing the pressure increases the
alcohol concentration in the reactor acid at equilibrium,
since for the four runs at pressures of 120, 100, 150, and
250 lbs./sq.in. respectively, the alcohol concentrations
in the acid at equilibrium were respectively 23.1%, 19.1%,
28.4%, and 29.1%.
It is interesting to compare the slope
of the acid concentration curves for the four runs with
(20)
in the batch run with
Lewis
Clay
by
the curve obtained
_litar
_
34.
60% sulfuric.
Lewis made one three-hour batch run at
100 0 C. and 500 lbs./sq.in. pressure, using 60% sulfuric
acid and pure propylene gas.
By analyzing liquid samples
from his reactor at intervals he was able to construct
a curve of alcohol concentration (both free and combined
as acid ester) versus time, which enabled him to study
the mechanism of batch absorption.
He plotted his data
as mols alcohol per mol acid versus time, and the fact
that his curve rose in a straight line for two hours led
him to believe that the rate of absorption did not depend
purely upon the difference between the instantaneous and
saturation concentrations of alcohol in the acid, since
in that case the rate of absorption would have fallen off
with time.
His conclusion was that the rate of absorption
was maintained by the solvent action of the alcohol already
formed.
His data have been replotted in Figure VIII as
per cent alcohol by weight in the reactor acid vs.
time,
and it is apparent that the straight line character of the
curve is still retained.
Comparing this curve with Figures IV, V, VI, and
VII it is seen that the curve for Run I is the only one
which may be said to reproduce the general shape of Lewis'
curve, and even in this case the per cent concentrations
at all points are lower.
In this respect all four runs are
similar, for in no case do the maximum alcohol concentrations
rise above thirty per cent, whereas Lewis' alcohol concentration leveled off at about 50%.
This high per cent is
-
_Z~Sr~ L
~wlla
35-.
undoubtedly due to the high pressure of 500 lbs./sq.in.
which he employed,
since the data of Runs III and IV
clearly indicate this effect of increasing the pressure.
The general shape of the curves for Runs II, III, and IV
is to follow a straight line for a certain length of time
and then rise to a greater slope before finally levelling
off to the horizontal.
This can be accounted for by Lewis'
explanation that the solvent action of the alcohol already
formed serves to increase the rate of propylene absorption.
Comparing the relative shapes of the vapor and
acid concentration curves for Runs III and IV, it is apparent that in each case as the rate of increase in alcohol
concentration in the acid begins to fall off the rate picks
up in the vapor phase.
This would seem to indicate that
generally speaking the mechanism of the flow absorption of
propylene in sulfuric acid is that the alcohol concentration in the acid continues to build up until a saturation
value is reached corresponding to the operating pressure
and temperature, and from then on alcohol is evolved from
the acid as fast as it is formed, the distribution between
alcohol, water vapor and propylene in the vapor leaving
the reactor being dictated by the equilibrium constant for
that temperature, the pressure, and the rate of gas flow.
Until the acid is saturated the tendency is for the alcohol
formed to remain in the acid rather than be evolved.
It
should be noted that in the case of the curves for alcohol
concentration in the condensed vapor samples there is a
5-,
M-S.NASS.AVVS.ChIRIrP
MASS.
-Aill
-
37.
lag involved, that is, the true instantaneous alcohol
concentrations are somewhat greater than the curves would
indicate due to the fact that the condensate samples
analyzed actually represent an accumulation over a period
of time.
The formation of organic byproducts was observed
in all of the runs.
In Runs III and IV byproduct formation
was greater than in Runs I and II, due to the higher temperature and pressures employed, although temperature probably is the controlling factor.
In Runs III and IV about
60 cc., of byproduct were found floating on the surface of
50 cc. of acid, while in Runs I and II the organic material
formed was only about 60% of the acid present.
Samples of
the organic layer were analyzed in a Babcock-Wilcox bottle,
using first a 10% salt solution to extract the alcohol, and
then, using a fresh sample of polymer, 85% phosphoric acid
was used to extract the ether.
The results of these analyses
indicated that the polymer contained about 3% alcohol and
80% ether.
Evidences of other side reactions observed dur-
ing the runs are the formation of cuprous oxide,
which
settled out as a red powder in the samples of condensed vapor
taken from the trap, the formation of cupric sulfate in the
reactor, and traces of sulfur dioxide in the gas stream.
Blackening of the copper lining in the reactor and of the
acid at the end of each run was caused by cupric oxide.
Apparently the copper lining will be oxidized by the hot
60% sulfuric acid if any air is present in the system.
........
0.6-CL
38.
If
no air at all
is
present thereshould be no oxidation of
the copper by the acid, as can be shown experimentally by
keeping a strip of copper wire submerged in hot 60% sulfuric
acid over a period of time.
As a commercial process this flow method of making
isopropyl alcohol by absorption in 60% sulfuric acid would
seem quite feasible, provided too high temperatures would
not have to be used to promote evolution of the alcohol from
the acid.
If this were the case losses would be incurred
due to polymerization and oxidation.
Neglecting loss of acid
due to deterioration, the major cost involved in the flow
process would be the rectification of the dilute (10%) alcoholwater mixture evolved from the reactor, while in the established method of making the alcohol not only is the cost of rectification involved but also the cost of reconcentrating the
sulfuric acid up to the requisite strength.
In the flow pro-
cess the acid would remain in the contact chamber continually,
water being added at intervals to replace that which was
evolved as alcohol.
In view of the comparatively meagre data obtained in
this thesis it would seem advisable that more work be done
along the same lines.
More data could be obtained at the
operating pressures of this thesis but using otiaer rates of
flow, and investigations should be made of the minimum temperatures at each flow rate necessary to produce steady
evolution of alcohol.
In order to reduce the side reactions
I
-n sssr~--i----ra
39.
to ether formation alone work should be carried out in
systems as completely free of air as possible.
If it proves
impracticable to remove all the air from the system, runs
might be made in glass-lined reactors.
_. _.
~...
.I~-..- .-. -
---- b- .~
40.
VII.
CONCLUSIONS
From the results of this thesis it
(1)
is
concluded that:
Alcohol concentrations approaching those theoretically
obtainable at equilibrium can be attained in a reasonably
short time in the production of isopropyl alcohol using
60% sulfuric acid in a flow process.
(2)
The optimum conditions of operation are 100-1600 C. and
150-250 lbs./sq.in. pressure, using propylene flow rates
of about 10.0 cu.ft./hr. per sq.in. of acid surface in
the contact chamber.
(3) Under these conditions alcohol concentrations from
15-18% may be expected within 4-5 hours in the condensate
from the vapor evolved from the acid in the reactor.
(4) At temperatures much above 1000 C. the formation aforganic
byproducts becomes appreciable,
-
41.
VIII.
RECOMMENDATIONS
It is recommended that:
(1)
More data be obtained on organic byproduct formation,
maximum alcohol concentrations obtainable, and the time required to reach dynamic equilibrium under varying conditions
of temperature, pressure, and rates of gas flow.
(2)
In particular, runs be made at 100 C. at faster gas
rates than those employed in Runs I and II of this thesis,
to see if any improvement is effected in the evolution of
alcohol from the reactor at this temperature.
~r~t~wa-ylk
IX.
APPENDIX
_
~ ri*Clrrr~--c~C
43.
A.
Calculation of Theoretical Per Cent
Conversion at Equilibrium vs. Temperature
for Different Pressures
(30)
The equation of Stanley, Youell, and Dymock is used
to solve for Kp at a given temperature.
Then, using the
fugacity chart for hydrocarbon vapors and the fugacities
(14)
for saturated water as calculated by Gunness from data
(48)
given in Keenan and Keyes Steam Tables, the per cent conversion obtainable at that temperature and a given pressure
may be calculated.
As an example, consider P =100 atmospheres and temp.=
2500 C.
1950
-T -6.060
loeo= -=
T = 250 4 -273= 523
loo Kk = 3.725 - 6.060 - 2.335
P = 0.00463
(1pr)i
Irv
(fV )
- , _ o ,,(
7Vu
rr
'
-,
-,
y
ih'
,
440
(7-d)prsp=
52.3
1.AAf ,
___
-
Po )
__z-
4
)
=/m
O.OO'?63)-
x(/ D00
0. A7/
v/"7/ =
(11 A&CX(V
~~ev
0 0 ,-/
7
Le
_
45.
B.
Calibration Plots and Calculations therefor
Data for flowmeter calibration (wet test meter):
temp.
of air
to flowmeter
static
pressure
flowmeter
differential
cu.ft./hr.
at inlet conditions
700 F.,
saturated with
water vapor.
o.28 in.Hg.
0.24 in.Hg.
4.15
0.51
0.42
5.65
(70F.21.20 C.)
0.83
0.69
7.65
0.88
0.74
8.12
1.21
0.97
9.4
1.53
1.24
10.8
1.98
1.64
12.4
These data are plotted on semilogarithmic paper as feet of
fluid flowing vs. volumetric rate of discharge at S.T.P.
As
an example in converting the test data, consider the first
set at
oh
=0
.24 in.
Hg:
p.p. water at 70 F. =0.74 in.Hg.
barom. =29.90 in.Hg.
.static pressure =0.28 in.Bg.
total pressure on gas to flowmeter '30.18 in.Hg.
to find the density of the gas,
first calculated:
% water:
"
18
% air (dry)
100
its
2.45
= 97.55
(.0245-18) - (.9755-28.9) = 28.5 =M.W.
molecular weight is
_
..
. -r~-~LeY
L'-L*rC
4&.
Let x = density of the saturated air in
30.18 )
x
(1) =
) (82)(294)
then
9.9
gm./cc.
whence x = 0.00120 gm./cc.
density of mercury =13.6 gm./cc.
h as feet of fluid flowing (i.e., feet of saturated air)
hence
-0.24
12
13.6
0.00120
z
2
volumetric rate of discharge at S.T.P. (1 atm., 320 F.):
273
(4.15)( 12M)
(3018)= 3.90 cu.ft./hr.
29.92
The other six points are calculated in this manner and
the calibration plot (Figure IX) drawn.
With the exception
of the first two points, at low discharge rates,
the other
points lie along a fairly straight line of slope about 0.53.
That is,
for the capillary tubing used the flow rate is
pro-
portional to the 0.53 power of the differential pressure.
The calibration plot constructed in this way can be used
for any fluid being metered (See Sample Calculations).
ix:
FIGU(1eE
Fh,-os-mrr4 e C4B,1aA77oM Pio r
200
/00
z
T
~
/0
kohA.oe-c, Psc4Ve (cu ftAr. ott T...
_
*--aaar*-sY"Ye-
48.
The calibration of the
Thermocouple Calibration:
copper-constantan thermocouple was done using the steam
point and the naphthalene point, the cold junction being
maintained at O"C. by ice water in a thermos bottle,
These two points provided sufficient calibration over
the temperature range at which the runs were being made.
Data
point
corrected temperature
mv.obstd stand.mv. deviation
steam
100.10 C.
3.975
4.280
naphthalene
218.26
9.680
10.265
-0.305 my.
-0.585
The temperatures corrected for barometric pressure,
which was 76.27 cm.Hg.,
steam: tp:1000.000
naphthalene: tp
217.96
were calculated from the equations:
O 03686(p-760)-0.0000202 (p-760)2
0.207(tp 273.1)lo J0
A calibration plot, Figure X, and a deviation plot,
Figure XI, were constructed for the thermocouple.
As explained under PROCEDURE, a 400 0 F. thermometer was
used more often than the thermocouple during the runs.
F-yv I
_rAERIVOCOMAME
)'17-
in&,ph-Maleftc
"Po'CAtt
U
~ 00
350
0
Fl-&vjew M.
-71Zre*mCOVPAS J4-V1A71(*V
nO7-
)*yh-A^kke
pot 19-1-
-ai
STC>E= .
C.
0
S-
M~ASS.
AVE-.
MA.SS.
4
If
Wu.
r/ cr e
N
Z
,P?E3soRe G;A&e
6 A
'I
A77aw fkar
t'70
.45
A
MH0 I
-3.0
-7-o
&0
r
/00
ISO,
.200
___
.24
300
34~
400
vea
£2sebveS Ca je Prssur
A*u
~~e
(As.4& .it.
w
?
1 1----~4-~52.
MOi
C.
Method of Analysis
The analytical procedure outlined below is the
(27)
(14)
modified Pondorf procedure used by Gunness. By this procedure
very small amounts of isopropanol can be detected, and it is
unaffected by the presence of dissolved propylene or acids.
Essentially the procedure consists in nitriting the alcohol
with nitrous acid, extracting the nitrite with three extractions of carbon tetrachloride, oxidizing it to nitric acid
and the alcohol with standard permanganate, and back titrating
the excess permanganate with standard thiosulfate.
In the
analysis only about 92% of the alcohol is determined, so it
is necessary to introduce a correction factor.
Two 150-cc. separatory funnels, a 250-cc. glassstoppered titrating bottle, pipettes, and burettes are used.
Into one funnel, hereafter called no, 1, are run 40 cc. of
carbon tetrachloride and 20 cc. of water, while 20 cc.
wash solution is put into funnel no. 2
(the wash solution
contains 50 gm. sodium bicarbonate and 4 gmin.
per liter of water).
of
sodium nitrite
Next, 3 cc. of a 30% sodium nitrite
solution is added to no. i, plus 5 cc. of 7.5% hydrochloric
acid, and then a one-cc. portion of the sample pipetted in.
Funnel no. I is shaken for 15 seconds, the carbon
tetrachloride layer is
30 seconds,
drawn off, and shaken in no. 2 for
and then drawn off into the glass-stoppered
titrating bottle.
53.
The extraction of the alcohol nitrite in no. I is
repeated,
using first
tetrachloride.
30 cc.
and then 20 cc.
of carbon
In each case the extraction is washed in
funnel no. 2 before being put into the titration bottle.
To the 90 cc. of carbon tetrachloride in the titrating
bottle is added 20 cc.
of one-normal sulfuric acid and 20 cc.
of the same acid containing 20% manganous sulfate.
Tenth-
normal permanganate is next added in 5 cc. portions until
further addition results in the persistence of brown manganese
oxide after shaking.
Potassium iodide solution is then added
and the solution titrated with standard thiosulfate until the.
rose color in the organic layer has disappeared.
The permanganate is standardized by running an analysis
on pure alcohol, and expressing the strength of one cc. of the
permanganate in terms of milligrams of alcohol.
be quite the theoretical amount.
is
This will not
Since the carbon tetrachloride,
not absolutely pure, it is necessary to run a blank on it,
which usually amounts to about 3% of the permanganate used.
(14)
The analysis is good to j%.
_
Z~L~rr~i~
ea
54.
D.
Sample Calculations
(Data of Run III)
In the standardization of the permanganate
(approx. 0.1 N) against alcohol of known concentration, 1 cc.
of isopropyl alcohol of sp.gr.
0.81 (87.2%) was diluted to
10 cc. with distilled water and one cc. taken for analysis.
1
mg. alcohol in sample = 0.81 K0. 8 72,*0 11000 =70.9 mg.
cc. KMn04 used = 23.6 cc.
blank on reagents used-0.30 cc. EEn04
thiosulfate for titration of excess MKn0 4 = 0.6 cc.
10 cc. KMn0 4 are equivalent to 11.9 cc. thiosulfate.
net ENnO4 used =23.6 - 0.3 -(r
strength of permanganate
70.9
2.8
)(0.6)
=
22.8 cc.
3,10
..
mg.Ealcohol per cc.
In the blank test on the reagents,
5 cc.
Kxn0 4 were
used and 5.6 cc. thiosulfate were required to titrate the10
excess.
Blank= 5.0 -(11i9) (5.6): 0.30 cc. KEWn0 .
4
In Run III, the average flowmeter differential was
0.61 in.Hg.
gas metered assumed to be pure dry propylene at 70"F.(21.2"C.)
static pressure
=
2.25 in.Hg.
total pressure -29.90- 2.25= 32.15 in.Hg.
Mol. wgt. of propylene =4 2 ,
- '~sC963~eL- ,-55.
density of gas being metered
32.15 ~88x294
ix42 )=
(29.92(
0
0.00188 gm./cc.
61
feet of fluid flowing=C2 (
13
6
O.00188)
0.00188
367
-
volumetric rate of flow at S.T.P. corresponding to f.f.f.-
367
= 4.76 cu.ft./hr.(S.T.P.)
4.76 x 454 x 1000 =6020 mmols/hr.
35
alcohol concentration in condensed vapor at equilibrium:
53.5 - 0.3 -(~-g)x
1
5.0 = 49.0 cc.
net XMn04 for
1 cc. sample.
1 cc.
KMn0 4 is equivalent to 3.10 mg. alcohol.
per cent alcohol in sample
49.0 x 3.10 x 100= 15.
1000
rate of alcohol production at equilibrium = 56.0 cc. per 15 minutes.
60
56.0 x
x 0.152 x 1000: 34000 mg. alcohol/hr.
34000 = 568 mmols alcohol/hr.
60
alcohol
molal ratio propylene in vapor from reactor at equilibrium:
568
6020-568
0.104
per cent conversion at equilibrium
568
6020 x 100: 9.40%
theoretical % conversion obtainable at 160 C. and 150 lbs./sq.in.
is 16.1% as read from Figure I-
-
U-
-~9-
-
-I
56.
E.
Summarized Data
Run I (May 2):
time begun ..............
9:00 P.M.
pressure................
120 lbs./sq.in. gage-
temp. of reactor ......
200 F.
flowmeter differential.. 1.5 in.Eg.
static pressure.........
0.2 in.Hg.
volume of 60% acid
charged...............
120 cc.
water replenished
in reactor at........ 11:15 P.M.
equilibrium conditions
reached at...........
and 12:30 A.M.
2:00 A.M.
Sampling:
location
time
reactor*:
10:30 P .M.
12:00
2:00 A.M.
4.0 cc.
4.0
4.0
10:30 P.M.
12:00
2:00 A.M.
4.0
1.0
2.0
11:00 P.M.
10.00
condenser
trap:
bubbler:
cc.taken
thiosulfate % alahol
cc.analyzed 1
0.5 cc.
0.5
0.5
19.4
33.2
39.5
CC.
1.3 cc.
1.1
2.5
11.2%
19.9
23.1
not analyzed
not analyzed
2.0 cc.
5.0
0.10
not analyzed, negligible acidity
and alcohol content.
Volumetric production of alcohol from trap at
equilibrium (cc. per 15 min.)................
Approximate concentration of reactor
acid at end of run ......................
Approximate volumes of acid and organic
layers in reactor at end of run....
negligible
.
..........
60%
80 cc.acid
50 cc.organic
* organic layer not included in reactor samples
in this case 3.0 cc. thiosulfate equivalent to 1.0 cc. KMnO
4
**
~IC~I~CY--LL
57.
Summarized Data (contd)
Run II (May 5):
time begun..............
4:00 P.M.
pressure ................
100 lbs./sq.in gage
temp. of reactor ........
220 F.
flowmeter differential..
1.0 in.Hg.
static pressure......... 2.5 in.Hg.
volume of 60% acid
charged ..............
150 cc.
water replenished
in reactor at......... 6:00 P.M. and 7:45 P.M.
equilibrium conditions
reached at ...........
8:40 P.M.
Sampling:
location
reactor:
condenser
trap:
bubbler:
time
5:20 P.M.
6:30
8:15
8:40
cc.taken cc.analyzed
10 cc.
12
10
10
IknO
4
thiosulfate % alcohol
0.5 cc. 10.7 cc.
0.5
19.8
34.5
0.5
0.5
34.7
2.4 cc.
4.4
5.0
4.3
5.2%
9.8
18.5
19.1
no samples large enough to analyze obtained at any
time during run.
not analyzed, negligible acidity and alcohol content.
Volumetric production of alcohol from trap at
equilibrium (cc. per 15 min.) ..................
negligible
Approximate concentration of reactor
acid at end of run.............................
60%
Approximate volumesof acid and organic
layers in reactor at end of run ................. 90 cc.acid
60 cc.organic
__ I__
58.
Summarized Data (contd)__
Run III (May 7):
time begun............
..
6:00 P.M.
pressure............... 150 lbs./sq.in. agea
temp. of reactor. ...... . 320" F.
flowmeter differential..
0.61 in. Hg.
static pressure ........ 2.25 in.Hg.
volume of 60% acid
charged..............
150 cc.
water replenished
in reactor at........
8:00 P.M.
equilibrium conditions
reached at ...........
11:45 P.M.
and 10:00 P.M.
Sampling:
location
time
reactor:
7:30 P.M.
9:30
11:15
11:45
condenser
trap:
7:30
9:30
1115
11:45
bubbler:
cc.taken cc.analyzed KMn04 thiosulfate % alcohol
10. cc.
10.
5.
5.
5.
5.
5.
5.
0.5 cc.
0.5
0.5
0.5
12.6
37.0
45.6
51.0
1.0
1.0
1.0
1.0
13.4
20.0
51.9
53.5
cc.
2.5 CC.
5.6
2.1
5.7
19.8
271.0
28.4
4.9
3.9
4.5
5. 0
2.8
5.1
14.8
15.2
6.3%
not analyzed, negligible acidity and alcohol content.
Volumetric production of alcohol from trap at
equilibrium (cc. per 15 min.).................... 56..0 cc.
Approximate concentration of reactor
acid at end of run ...............................
Approximate volumes of acid and organic
layers in reactor at end of run .................
60%
.
50 cc.acid
60 cc.organic
-- r
- --~-~---C-
59.
Summarized Data (contd)
Run IV (May 9):
time begun..............
4:30 P.M.
pressure ...........
.....
250 lbs./sq.in. gage.
temp. of reactor........ 320 F.
flowmeter differential.. 0.84 in.Hg.
static pressure......... 2.25 in.Hg.
volume of 60% acid
charged...............
120 cc.
water replenished
in reactor at......... 6:30 P.M. and &:45 P.M.
equilibrium conditions
reached at........
8:40 P.M.
Sampling:
location
time
reactor:
6:00 P.M.
7:10
8:15
8:40
condenser
trap:
6:00
7:10
8:15
8:40
bubbler:
cc.taken cc.analyzed 1(04 thiosulfate % alcohol
5.0 cc.
5.0
5.0
5.0
0.5 cc.
0.5
0.5
0.5
12.0
36.7
50.3
51.1
5.0
5.0
5.0
5.0
1.0
1.0
1.0
1.0
14.5
23.5
59.1
62.3
cc.
4.5 cc
5.0
6.2
4.5
5.0
3.8
4.5
3.8
4.9%
20.1
27..7
29.1
3.1
6.2
17:.1
18.2
not analyzed, negligible acidity and alcohol content.
Volumetric production of alcohol from trap at
equilibrium (cc. per 15 min.)................
64.0 cc.
Approximate concentration of acid in reactor
at end of run ................................... 60%
Approximate volumes of acid and organic
layers in reactor at end of run.................09
50 cc.acid
60 cc.organic
..sP~BU)I~'~ur
60.
F.
Details of Apparatus
The details of the apparatus shown in Figure II are
given in this section.
Pressure was supplied to the system by an air compressor driven by a 1/3 H.P. motor, and the compressor was
capable of maintaining a pressure of 450 lbs./sq.in. gage
on the system, which was well above the desired operating
pressure of 100-250 lbs./sq.in.
The reactor was a copper-lined length of heavy iron
pipe, about thirty inches long and 3/4 in. inside diameter,
and the head pieces were bolted to flanges at each end.
At
the bottom of the reactor a brass needle valve was located
for the purpose of withdrawing samples of the acid from the
reactor during a run and also to provide a means of charging
the reactor with acid prior to a run.
The outlet tube from
the bottom of the reactor to this valve was 3/16 in. medium
wall copper tubing, as was most of the other tubing in the
apparatus, and extended up into the reactor about three inches.
It was silver-soldered through a thin disc of copper, and
the copper disc held tightly against the bottom flange in
the manner of a gasket between the bottom head and flange,
so that only copper and silver solder were exposed to the
acid in the reactor.
The use of a needle valve for with-
drawing the acid samples was considered advisable in view
of the better control obtained under pressure.
The propy-
lene entered at the top of the reactor through quarter-inch
.. \r--g U*--c~-~61
copper tubing which extended dovwn into the reactor to a
point about six inches from the bottom.
Since the reactor
was usually charged with acid to a depth of about 17-18
inches, the depth of acid through which the propylene bubbled,
was probably in
the neighborhood of 14 inches,
allowing a few
inches for the velocity of the gas as it issued from the inlet pipe.
A thermocouple well also extended down from the
top to a point about eight inches from the bottom of the reactor, that is, to a point about at the middle of the column
of acid.
This well consisted of quarter-inch medium wall
copper tubing, plugged at the end, and in most of the runs it
was found more convenient to use a thermometer in place of a
thermocouple, the readings differing by only about three degreesFahrenheit.
The thermocouple originally constructed for measur-
ing the temperature was copper-constantan, and its calibration
and deviation plots are given in the APPENDIX.
At the top of
the reactor an outlet was provided for the gas through quarterinch copper tubing, extending into the reactor about i/8 inch.
The three tubes at the top of the reactor were silver-eSldered
through a copper disc as in the case of the sampling line at
the bottom of the reactor.
Heat was supplied to the reactor by
means of i/16 inch Chromel-A heating ribbon supplied with
110-volt D.C. current, which also ran the motor for the compressor.
A variable resistance afforded a means of varying,
the temperature in the reactor.
Asbestos tape was used to
insulate the reactor and the vapor line leading to the condenser.
___n_ ~
62.
The condenser was a spiral of quarter-inch copper
tubing in a vertical position, the cooling medium being water.
Soldered to the line leading from the condenser was a trap
for catching the condensate, and from which samples of the
condensate could be withdrawn through a brass needle valve.
The trap consisted of a piece of 3/4 inch brass pipe about
eleven inches long, a vapor exit being provided from the top
of the trap about one inch above the bottom of the inlet
tube from the condenser.
The gas from the trap then passed:
through a brass needle valve which controlled the pressure
on the reactor, and then through a water bubbler and calcium
chloride drying tube to the flowmeter,
cycled to the compressor.
from which it
was re-
The water bubbler was similar in
construction to the trap, except that a pinchclamp could be,
used instead of a needle valve for withdrawing liquid samples.
since the bubbler was on the low pressure side of the system.
The inlet tube to the bubbler extended dovwn to about an inch
from the bottom, and ordinarily about six inches of water
were maintained in the bubbler.
The flowmeter used to measure the rate of flow of
propylene to the reactor consisted of a piece of capillary
glass tubing about three inches long with a capillary between
1/32 and 1/64 inch in diameter, connected at each end to a
U-tube of mercury and one of water (for low flow rates), and
connected at the inlet end to a static pressure gauge containing mercury.
A thermometer well was not installed as it
-m
6Z.
was- felt that the gas being metered would be substantially
at room temperature.
The meter was calibrated in the standard
way with a wet test meter, and the calibration curve is given
in the APPENDIX.
The pressure gauge located just before the reactor
was of the Bourdon type, and had a range of 750 lbs./sq.in.
It was calibrated by the dead weight method, and the calibration
is given as a deviation plot in the APPENDIX,
The unit for supplying make-up propylene to the system
consisted of a five-gallon bottle connected to the aylinder
of propylene, the gas in the bottle being maintained under a
constant head of four feet of water.
A connecting line enter-
ed the system at a point before the flowmeter, the storage
propylene being dried by the same calcium chloride drier as
was the gas coming through the bubbler.
The four foot head
of water was necessary due to the fact that at the operating
pressures and flow rates desired it was necessary to crack the,
control valve so far that an appreciable static pressure
existed on the low pressure side of the system.
I
-
64.
G.
(1)
Bibliography
Bernstein and McCoy, "The Hydration of Propylene
under High Pressure;v
(2)
ki.I.T.
thesis, 1932.
Bowen and Nash, "Petroleum and Petroleum Gases as
Chemical Raw Materials;Q
Manufacturer,
Refiner and Natural Gas
vol. 12, no.
9,
pp.
361-69.
(3)
Bowles,
(4)
Brezinske,
(5)
British Patent 238,900.
(6)
Brooks, "The Manufacture of Alcohols and Related
M.IT.
thesis, 1933.
M.I.T.
thesis, 1929.
Products from Petroleum;" Refiner, vol. 12,
no.
(7)
9, pp. 353-60.
Brooks and Humphrey, "The Action of Concentrated
Sulfuric Acid on Olefins;"
Journal of the
American Chemical Society, May 1918.
(8)
Cassar,
19,
Industrial and lgineering Chemistry,
1061 (1927).
(9)
Ellis, "Chemistry of Petroleum Derivatives;" chapter 11.
(10)
Francis,
(11)
Frolich,
Ind. Eng.
Chem.,
20, 283 (1928).
"Chemical Trends in the Petroleum Industry;"
Ind. Eng. Chem., August 1938.
(12)
Gerr, Pipik, and Mezhebovskaya, "Preparation of Ethyl
Alcohol from Ethylene;v Refiner, vol.12,no.2,p.70.
(13)
Gilliland, Gunness, and Bowles, "Free Energy of Ethylene.
Hydration;"
Ind.
Eng.
Chem.,
March 1936.
65.
Bibliography (cont Id)
(14)
Gunness, vFree Energy of Hydration of Ethylene;v
M.I.T.
(15)
thesis, 1934.
Holden and Stackpole, rThe Hydration of Propylene to
Form Isopropyl Alcohol and Ether;n M.I.T.
thesis, 1934.
(16)
International Critical Tables, vol. III, p. 239.
(17)
Johannsen and Gross, U.S.P. 1,607,459 (1927).
(18)
Keenan and Keyes,nThermodynamic Properties of Steam,.
Wiley and Sons (1936).
(19)
Klever and Glaser, Mitt. Chem. Tech. Inst. Tech. Hochschule,
Karlsruhe, I.
(20)
pl. 1-10, 1923.
Lewis, "Absorption of Propylene in Sulfuric Acid,r
M.I.T. thesis, 1937.
(21)
Lewis and Kay, Oil and Gas Journal, March 29, 1934.
(22)
Lewis and Luke, Ind. Eng.
(23)
Lewis and Randall, "Thermodrnamics;u
(24)
Majewski, "The Hydration of Propylene under High Pressure;u
i,.I.T.
Chem., 25, 725 (1933).
McGraw-Hill (1923).
thesis, 1934.
(25) Marck and Flege, "Catalytic Vapor-Phase Hydration of
2-Butene under High Pressure;"'Ind.Eng.Chem.,S
,1428
(1932).
(26.)
Millard,
"Physical Chemistry for Colleges.#
(27)
Pondorf:
Zeit, fur Analyt. Chim., 80,
(28)
Rand and Williams,,Catalytic Hydration of Propylene under
High Pressure;" M.I.T.
(29)
401 (1930).
thesis, 1933.
Sanders and Dodge, Ind. Eng.
Chem., 26, 208-13 (1934).
66.
Bibliography (cont td)
(30)
Stanley, Youell, and Dymock, "The Equilibrium Constants of
the Vapor-Phase Hydration of Ethylene, Propylene, and the
Butylenes;'
Journal of the Society of Chemical Industry,
Transactions, 53, 205 T (1934).
(31)
Swann, Snow, and Keyes;
(32)
Walker, Lewis, McAdams, and Gilliland, "Principles of
Chemical Engineering.
Ind. Eng. Chem., 22, 1048 (1930).